Quantcast

The Hype Factor: xScore

Eno Sarris, March 04, 2015 -   

What is the hype factor? Why do certain beers seem destined for four-star ratings before anyone has had a sip? 

This is an impossible question to answer definitively. Especially if you want super high standards for your mathematics. There will be some fuzzy math in this post. 

But let's say that I could ask for some limited information about your beer, and then predict with some level of accuracy the rating that you will put on said beer. Say an xScore, or an expected score. Wouldn't this start to get at the Hype Factor? 

Specifics. If you tell me the brewery name, the relative distribution of said brewery, the style of your beer, and the alcohol by volume in your beer, I can tell you the expected score of that beer. In other words, we can relate Brewery BAR, Brewery Number of Beers, Style, and ABV to ratings.

We did. We did do that, and we got an r-squared of .55, meaning that our equation explains 55% of the variance in beer scores. 

xScore = .8924 + (0.0003 * Brewery BAR) + (0.716  * Style Replacement Level) + (0.0322 * ABV)

How does this relate to the hype factor? Say you've got a Hill Farmstead Double IPA on your hands, with a 9% ABV. The expected score on that beer is 3.77 stars. If Hill Farmstead made a DIPA that averaged 3.6 stars, it would be a (relative) dud. It would have failed the hype factor. It would have underperformed the conditions set by the name and ABV on the label and the size and reach of the brewery. 

There are problems with this analysis. The outcome is itself a part of Brewery BAR and Style Replacement Level, so there are mutlicollinearity issues. We are correlating ratings to themselves. But we've weighted those ratings, and it's not the ratings of the specific beer at play, it's the style's ratings, and the brewery's ratings, that we've used to predict the single beer. 

Plus, we're just having a little bit of fun.

Which beers underperformed their hype factor the most? Which beers had the biggest negative gap between their expected score and their actual score? Here's the leaderboard with breweries with at least 35 beers.

Beer Brewery Style Avg Rating xScore xScore-Avg Brewery BAR Brewery Beers
Pabst Blue Ribbon Dry Pabst Brewing Company American Strong Ale 2.29 3.74 1.454 13.9 41
Tuborg Lime Cut Ringnes Bryggeri (Carlsberg) Fruit Beer 2.05 3.41 1.358 -14.4 37
Black Hole Barrel Aged Scotch Mikkeller American Imperial / Double Stout 2.96 4.30 1.337 451.6 309
Storm BrewDog Imperial / Double IPA 2.68 3.94 1.268 211.2 162
Keystone Ice Coors Brewing Company American Adjunct Lager 1.78 3.03 1.244 -69.4 48
Grodziskie Bierbrouwerij Emelisse Gose 2.35 3.59 1.239 33.9 38
Speedway Stout (Lapsang Tea) AleSmith Brewing Company American Imperial / Double Stout 2.96 4.19 1.232 206.8 58
Pripps Carlsberg Sverige (Carlsberg) Black Lager 2.10 3.32 1.223 -80.1 45
Bud Light Golden Wheat Anheuser-Busch (ABInBev) American Pale Wheat Ale 2.16 3.38 1.212 -128.1 88
Roble Blanco (2014) The Bruery Sour Ale 3.12 4.29 1.171 723.8 208
Keystone Original Coors Brewing Company American Adjunct Lager 1.80 2.97 1.169 -69.4 48
Natural Ice Anheuser-Busch (ABInBev) American Light Lager 1.79 2.94 1.15 -128.1 88
Bourbon Barrel Aged Token Alpine Beer Company (CA) American Porter 2.67 3.81 1.137 153.0 35
Idiot's Drool Weyerbacher Brewing Company Old Ale 2.77 3.89 1.119 54.9 62
Punishment Stone Brewing Co. American Strong Ale 3.04 4.15 1.108 764.4 255
Carlsberg (Sverige) 2.8% Carlsberg Sverige (Carlsberg) Pale Lager 1.86 2.96 1.099 -80.1 45
Black & Red Dogfish Head Craft Brewery American Imperial / Double Stout 3.12 4.21 1.091 422.4 138
Black Ink & Blood Mikkeller Russian Imperial Stout 3.37 4.42 1.05 451.6 309
Skol Lager Carlsberg Group Euro Lager 1.96 2.97 1.015 -25.7 43
Spontandoublekiwi Mikkeller Sour Ale 2.96 3.97 1.013 451.6 309

If this expected score was about anything, it was about Mikkeller's Black Hole Barrel Aged Scotch. It's barrel aged, it's from Mikkeller, it's a double stout -- it's supposed to be good. And an average three star rating, though it sounds like it's 'average,' is not good. Mikkeller has failed to live up on the hype factor on this beer. 

The surprise, personally, might be AleSmith's Lapsang Tea Speedway Stout. Their Vietnamese Coffee variant on that Speedway is one of my personal favorites of all time. I have not had the Lapsang Tea version. Perhaps it's bad. It doesn't *sound* like a great idea. 

And oh boy, The Bruery made a bad sour? Kinda want to try it now. 

On the flip side of the coin are the beers that had higher scores than their xScore. 

Beer Brewery Style Average Score xScore xScore-Ave Brewery BAR Brewery Beers
Boh Ice Pabst Brewing Company American Adjunct Lager 4.30 3.05 1.25 13.9 41
Lou Pepe Kriek (2011) Brasserie Cantillon Fruit Beer 4.59 3.58 1.02 472.3 79
Gluten Freek Wicked Weed Brewing Gluten-Free 4.03 3.04 0.99 175.9 79
Zombie Dust Three Floyds Brewing Company American Pale Ale 4.66 3.70 0.96 425.8 114
Blueberry Lambic (2007) Brasserie Cantillon Lambic 4.78 3.83 0.95 472.3 79
Fou' Foune (2012) Brasserie Cantillon Lambic 4.76 3.83 0.93 472.3 79
Sour Fruit New Glarus Brewing Company Sour Ale 4.75 3.81 0.93 218.4 54
Apple Brandy Barrel Aged Mexican Cake Westbrook Brewing Co. Stout 4.76 3.83 0.92 135.3 48
Lou Pepe Kriek (2007) Brasserie Cantillon Lambic 4.74 3.83 0.91 472.3 79
Wisconsin Belgian Red New Glarus Brewing Company Fruit Beer 4.36 3.47 0.90 218.4 54
Wisconsin Belgian Red (2011) New Glarus Brewing Company Fruit Beer 4.35 3.47 0.89 218.4 54
Guest Lager Half Acre Beer Company Pale Lager 4.08 3.20 0.88 145.8 57
R & D Very Sour Blackberry New Glarus Brewing Company Sour Ale 4.70 3.81 0.88 218.4 54
Lou Pepe Kriek (2010) Brasserie Cantillon Lambic 4.70 3.83 0.87 472.3 79
Zombie Dust (2014) Three Floyds Brewing Company American Pale Ale 4.57 3.70 0.87 425.8 114
Lou Pepe Framboise (2010) Brasserie Cantillon Lambic 4.69 3.83 0.86 472.3 79
Nelson Alpine Beer Company (CA) American IPA 4.57 3.71 0.86 153.0 35
Kiwi Rising Jack's Abby Brewing Pale Lager 4.08 3.22 0.85 175.5 50
Fou' Foune (2014) Brasserie Cantillon Lambic 4.68 3.83 0.85 472.3 79
Hoponius Union Jack's Abby Brewing Pale Lager 4.00 3.16 0.84 175.5 50

What's up with sours? They have a crazy high average rating, but they seem to screw with xScore. I mean, if xScore really got at hype factor, there's no way it would call Cantillon Fou' Foune an underrated beer, would it? People do crazy things to get Cantillon beers. They fly to Europe, for one. 

There's definitely a style-busting thing going on here. Boh Ice, Hoponius Union, Guest Lager -- these beers dominate in a sytle that doesn't produce hype. Is it cheating? We'll continue to bang our heads against that question. 

And then you have two of the best IPAs out there in Zombie Dust and Nelson. You'd think an IPA from a mid-sized brewery like Three Floyds would have a good xScore, but it's not an IPA -- it's listed as an American Pale Ale, and *that* replacement level is much lower. Nelson? Comes from a small brewery, and we used brewery BAR as a substitute for distribution. 

Should we have? What would be more predictive of a beer's success relative to its brewery? Size and distribution, or average rating per beer? I'm guessing it would be average rating -- Cantillon -- over size -- The Bruery. So maybe we should re-run xScore with Brewery BAR/beer in for Brewery BAR? 

Yes, it's now getting a little ridiculous -- we'd be predicting how good this beer should be by how good the average beer from that brewery is, and what style it is, but still... that's kind of cool if you ask me. We'll have to re-run xScore for 2.0 next week to see how that changes things. 

For now, revel in the fact that The Bruery made a bad sour and Cantillon surpasses even its own hype. 

comments powered by Disqus